Recently, I purchased a FEZ Panda and I've done some pretty cool stuff with it. Well, cool to me anyway. I recently had some ideas for things to do with LEDs and wanted to share them as kits. I see a lot of such kits utilizing small, cheap processors like the ATTiny13 or ATTiny85. Makes sense if it's just doing one thing with a small circuit that's simple. But I'm now hooked on using the .Net Micro Framework and love the fact that I can do everything from Visual Studio. For these little kits, I'm not opposed to using an AVR chip and writing in C, but was wondering if there was a .Net-enabled alternative. I've found some ARM7/ARM9 processors for around $5, but their package is rather difficult to solder onto a PCB for something as basic as a learning kit. Arent there any other form factors or other cheap alternatives that can be programmed using .NET MF?

EDIT: I'll also point out that there exist AVRs that come preloaded with the Arduino bootloader and they're cheap. These are also good chips to program with a single purpose and ship them with kits. I can't find a .Net MF enabled chip that I can do the same with.

Thanks in advance...

  • \$\begingroup\$ The Panda itself is based on an ARM SOC, and is available as a loose chip, preloaded with the framework. But it is a fine pitch package and requires really good soldermask and a stead hand to hand-solder. \$\endgroup\$ – RBerteig Apr 6 '11 at 1:41
  • \$\begingroup\$ Interesting. I didnt know you could get that chip preloaded with the framework. To program AVRs, you generally need a seperate hardware programmer, although I've seen Arduinos used for this as well. Is it the same with the ARM SOC? \$\endgroup\$ – Kilhoffer Apr 6 '11 at 2:30
  • \$\begingroup\$ Its available from the GHI (makers of the Panda) as the USBizi100 or USBizi144 depending on pin count and functions. It's an LPC2387 or LPC3288 ARM7 chip from NXP with internal 512KB FLASH and 96KB SRAM and lots of peripherals, preloaded with the .NET micro framework. Its often possible to order chips like this pre-programmed. The distributor will use a DataIO or similar device to do the job on the bare package. \$\endgroup\$ – RBerteig Apr 6 '11 at 6:59
  • 1
    \$\begingroup\$ Anything big enough for the .NET micro framework is going to be about as complex as the USBizi100, and tend to have lots of fine pins. You might look at the 32-bit PIC family, some of which is available with lower pin counts. There's also the CORTEX M3 devices such as the STM32F10xxx devices from STMicro. At the really low end they run just a few dollars each, and are nice little ARM cores inside. But still have a lot of pins. \$\endgroup\$ – RBerteig Apr 6 '11 at 7:03

No. The price of a microcontroller running the .NET framework will always be greater than the price for a microcontroller with equivalent performance (or equivalent to the price of a micro with greater performance) which is programmed in assembly, C, or even using the Arduino tools.

When buying a microcontroller for the .NET framework, you pay for small transistors, you pay for more transistors, you pay for complexity, you probably pay for licensing a core developed somewhere else, and, since you're paying for all that, most manufacturers imagine that you'll want to pay for higher pin-count packages since your die budget supports them. If you did this in assembly or C, you could get equivalent performance and pay a lot less.

See also the following questions:

The first question in the list is very applicable to this question, it just discusses the suitability of the .NET micro framework as opposed to the pricing and programming which we're discussing here.

If you still want to use .NET, you think you can get acceptable performance out and you're going to be buying a significant number of parts (at least 1k, probably easier at 10k), you can do the same thing that the vendors of the FEZ parts do: Have it loaded at the factory. It costs somewhere in the region of a quarter apiece to program large files like the one you'll be working with.

  • 1
    \$\begingroup\$ Thanks for these links. Unless someone can show me otherwise, I've concluded that .Net enabled chips are likely overkill for small, single purpose kits. It seems I'm better off just programming some AVRs for this purpose and saving .Net MF for larger goals. \$\endgroup\$ – Kilhoffer Apr 6 '11 at 15:38
  • \$\begingroup\$ @Kilh other than the fun of programming it in .Net, what gain would you actually see for single purpose chips? Unless of course, that single purpose is something very .Net geared. For instance, string parsing in C is a huge pain, but in .Net it's not(sorta). So for something like that it may be worth it. \$\endgroup\$ – Earlz Apr 6 '11 at 21:58
  • \$\begingroup\$ I guess I wouldnt gain much. A single purpose chip that simply coordinates the lighting of several LEDs in a sequence doesnt have to be done in .Net and would likely be cheaper and easier for hobbyists to solder than an ARM7 with a surface mount package. This is the route I've decided to take, although I do enjoy programming in .Net and getting full debugger support. I was really hoping for a cheap/easy chip that I could just program in .Net but it sounds like there is a significant trade off. \$\endgroup\$ – Kilhoffer Apr 6 '11 at 23:35

Why are you wanting to use something other than the Panda? If you need more power, you might have a look at the other .Net FEZ line of products: http://www.ghielectronics.com/catalog/category/7/

  • 1
    \$\begingroup\$ I think he wants to make his kit as accessible to the public as possible, so something as pricey as the Panda doesn't make a lot of sense. \$\endgroup\$ – Dave Apr 6 '11 at 4:17
  • \$\begingroup\$ Exactly. This is for a kit and the chip shouldnt be doing much at all, so I don't need all the extra power. I'm looking for a controller thats $5 or less that I can include in the kit so the user can solder it (and other components) onto the PCB and just go. So Dave, you're exactly right. The Panda makes no sense here. \$\endgroup\$ – Kilhoffer Apr 6 '11 at 5:24
  • 1
    \$\begingroup\$ The bare chip runs just under US$20 for single units. And its a 100 or 144 pin LQFP which is a bit of a bear to solder down by hand without all the right tools. \$\endgroup\$ – RBerteig Apr 6 '11 at 7:01
  • \$\begingroup\$ @Kilh I think you're going to have trouble finding anything that can run .Net Micro without requiring surface mount soldering of some sort. \$\endgroup\$ – Earlz Apr 6 '11 at 21:55
  • \$\begingroup\$ Agreed. \$\endgroup\$ – Kilhoffer Apr 6 '11 at 23:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.